woe 0.1.4

悲哀

版本：0.1.4

woe转换工具主要用于信用评级的记分卡模型

$ pip install woe

$ pip install git+https://github.com/boredbird/woe

功能

• Split tree with IV criterion
• Rich and plentiful model eval methods
• Unified format and easy for output
• Storage of IV tree for follow-up use

woe模块功能树

|- __init__
|- config.py
|   |-- config
|       |-- __init__
|               |-- change_config_var_dtype()
|               |-- load_file()
|- eval.py
|   |-- compute_ks()
|   |-- eval_data_summary()
|   |-- eval_feature_detail()
|   |-- eval_feature_stability()
|   |-- eval_feature_summary()
|   |-- eval_model_stability()
|   |-- eval_model_summary()
|   |-- eval_segment_metrics()
|   |-- plot_ks()
|   |-- proc_cor_eval()
|   |-- proc_validation()
|   |-- wald_test()
|- feature_process.py
|   |-- binning_data_split()
|   |-- calculate_iv_split()
|   |-- calulate_iv()
|   |-- change_feature_dtype()
|   |-- check_point()
|   |-- fillna()
|   |-- format_iv_split()
|   |-- proc_woe_continuous()
|   |-- proc_woe_discrete()
|   |-- process_train_woe()
|   |-- process_woe_trans()
|   |-- search()
|   |-- woe_trans()
|- ftrl.py
|   |-- FTRL()
|   |-- LR()
|- GridSearch.py
|   |-- fit_single_lr()
|   |-- grid_search_lr_c()
|   |-- grid_search_lr_c_main()
|   |-- grid_search_lr_validation()

示例

在examples目录中，有一个简单的woe转换程序作为教程。

或者你可以用这个包编写一个更复杂的程序。

版本记录

灾难0.1.4 2018-03-01

• 支持PY3

悲哀0.1.3 2018-02-09

• woe.feature_process.proc_woe_discrete(): fix bug when deal with discrete varibales
• woe.eval.eval_feature_detail(): fix bug : utf-8 output file format
• woe.GridSearch.grid_search_lr_c_main(): add function warper for convenience and high efficiency
• woe.GridSearch.grid_search_lr_c_validation(): monitor the ks performance of training sets and test sets on different ‘c’
• supplement examples test scripts

灾难0.1.2 2017-12-05

• woe.ftrl.FTRL(): add online learning module

悲哀0.1.1 2017-11-28

• woe.config.load_file(): change param data_path to be optional
• woe.eval.eval_feature_stability(): fix bug : psi_dict[‘stability_index’] computation error
• woe.feature_process.change_feature_dtype(): add friendly tips when encounter a error
• woe.feature_process.calulate_iv(): refactor the code
• woe.feature_process.calculate_iv_split(): refactor the code
• woe.feature_process.binning_data_split(): reduce the number of len() function calls with __len__() and shape attributes;replace namedtuple with dict
• woe.feature_process.fillna(): new added function to fill null value
• woe.GridSearch.grid_search_lr_c(): list of regularization parameter c specified inside the function is changed to the user specified

悲哀0.0.9 2017-11-21

• Add module : GridSearch for the search of optimal hyper parametric C in LogisticRegression
• Code refactoring: function compute_ks and plot_ks

悲哀0.0.8 2017-09-28

• More flexible: cancel conditional restriction in function feature_process.change_feature_dtype()
• Fix bug: the wrong use of deepcopy in function feature_process.woe_trans()

灾难0.0.7 2017-09-19

• Fix bug: eval.eval_feature_detail raises ValueError(‘arrays must all be same length’)
• Add parameter interface: alpha specified step learning rate ,default 0.01